2 1 requirements of voice

7/29/2019 2 1 Requirements of Voice

1/10

Requirements of Voice in an IPInternetwork

Real-Time Voice in a Best-Effort IP InternetworkThis topic lists problems associated with implementation of real-time voice traffic in a

best-effort IP internetwork.

3 2005 Cisco Systems, Inc. All rights reserved. Cisco PublicIP Telephony

IP Internetwork

IP is connectionless.

IP provides multiple paths from source todestination.

The traditional telephony network was originally designed to carry voice. The design of circuit-

switched calls provides a guaranteed path and a delay threshold between source and

destination. The IP network was originally designed to carry data. Data networks were not

designed to carry voice traffic. Although data traffic is best-effort traffic and can withstand

some amount of delay, jitter, and loss, voice traffic is real-time traffic that requires a certain

quality of service (QoS). In the absence of any special QoS parameters, a voice packet is

treated as just another data packet.

The user must have a well-engineered network, end to end, when running delay-sensitive

applications such as VoIP. Fine-tuning the network to adequately support VoIP involves a

series of protocols and features geared toward QoS.

Example: Real-Time Voice Delivery Issues

In the IP network shown in the figure, voice packets that enter the network at a constant rate

can reach the intended destination by a number of routes. Because each of these routes may

have different delay characteristics, the arrival rate of the packets may vary. This condition is

called jitter.

Copyright 2005, Cisco Systems, Inc. Introduction to VoIP > Requirements of Voice in an IP Internetwork 2-3


2/10

2-4 Cisco Networking Academy Program: IP Telephony v1.0 Copyright 2005, Cisco Systems, Inc.

Another effect of multiple routes is that voice packets can arrive out of order. The far-end

voice-enabled router or gateway has to re-sort the packets and adjust the interpacket interval for

a proper-sounding voice playout.

Network transmission adds corruptive effects like noise, delay, echo, jitter, and packet loss to

the speech signal. VoIP is susceptible to these network behaviors, which can degrade the voice

application.

If a VoIP network is to provide the same quality that users have come to expect from traditional

telephony services, then the network must ensure that the delay in transmitting a voice packet

across the network, and the associated jitter, does not exceed specific thresholds.


3/10

Packet Loss, Delay, and JitterThis topic discusses the causes of packet loss, end-to-end delay, and jitter delay in an

IP internetwork.


Packet Loss, Delay, and Jitter

Packet loss

Loss of packets severely degrades the voice application.

Delay

VoIP typically tolerates delays up to 150 ms before thequality of the call degrades.

Jitter

Instantaneous buffer use causes delay variation in thesame voice stream.

In traditional telephony networks, voice has a guaranteed delay across the network by strict

bandwidth association with each voice stream. Configuring voice in a data network

environment requires network services with low delay, minimal jitter, and minimal packet loss.

Over the long term, packet loss, delay, and jitter will all affect voice quality, as follows:

Packet loss: The IP network may drop voice packets if the network quality is poor, if the

network is congested, or if there is too much variable delay in the network. Codec

algorithms can correct small amounts of loss, but too much loss can cause voice clipping

and skips. The chief cause of packet loss is network congestion.

Delay: End-to-end delay is the time that it takes the sending endpoint to send the packet to

the receiving endpoint. End-to-end delay consists of the following two components:

Fixed network delay: You should examine fixed network delay during the initialdesign of the VoIP network. The International Telecommunication Union (ITU)

standard G.114 states that a one-way delay budget of 150 ms is acceptable for high-

quality voice. Research at Cisco Systems has shown that there is a negligible

difference in voice quality scores using networks built with 200-ms delay budgets.

Examples of fixed network delay include propagation delay of signals between the

sending and receiving endpoints, voice encoding delay, and voice packetization time

for various VoIP codecs.



4/10


Variable network delay: Congested egress queues and serialization delays onnetwork interfaces can cause variable packet delays. Serialization delay is a constant

function of link speed and packet size. The larger the packet and the slower the link-

clocking speed, the greater the serialization delay. Although this ratio is known, it

can be considered variable because a larger data packet can enter the egress queue at

any time before a voice packet. If the voice packet must wait for the data packet to

serialize, the delay incurred by the voice packet is its own serialization delay, plus

the serialization delay of the data packet in front of it.

Jitter: Jitter is the variation between the expected arrival of a packet and when it is actuallyreceived. To compensate for these delay variations between voice packets in a

conversation, VoIP endpoints use jitter buffers to turn the delay variations into a constant

value so that voice can be played out smoothly. Buffers can fill instantaneously, however,

because network congestion can be encountered at any time within a network. This

instantaneous buffer use can lead to a difference in delay times between packets in the

same voice stream.

Example: Packet Loss, Delay, and Jitter Problems

The effect of end-to-end packet loss, delay, and jitter can be heard as follows:

The calling party says, Good morning, how are you?

With end-to-end delay, the called party hears, Good morning, how are you?

With jitter, the called party hears, Goodmorning, howare you?

With packet loss, the called party hears, Good m..ning, w are you?


5/10

Consistent ThroughputThis topic describes the methods that you can use to ensure consistent delivery and throughput

of voice packets in an IP internetwork.


Consistent Throughput

Throughput is the amount of data transmittedbetween two nodes in a given period.

Throughput is a function of bandwidth, errorperformance, congestion, and other factors.

Tools for enhanced voice throughput include:

Queuing

Congestion avoidance

Header compressionRSVP

Fragmentation

Throughput is the actual amount of useful data that is transmitted from a source to a

destination. The amount of data that is placed in the pipe at the originating end is not

necessarily the same amount of data that comes out at the destination. The data stream may be

affected by error conditions in the network; for example, bits may be corrupted in transit,

leaving the packet unusable. Packets may also be dropped during times of congestion,potentially forcing a retransmit, using twice the amount of bandwidth for that packet.

In the traditional telephony network, voice had guaranteed bandwidth associated with each

voice stream. Cisco IOS software uses a number of techniques to reliably deliver real-time

voice traffic across the modern data network. These techniques, which all work together to

ensure consistent delivery and throughput of voice packets, include the following:

Queuing: The act of holding packets so that they can be handled with a specific priority

when leaving the router interface. Queuing enables routers and switches to handle bursts of

traffic, measure network congestion, prioritize traffic, and allocate bandwidth. Cisco

routers offer several different queuing mechanisms that can be implemented based on

traffic requirements. Low Latency Queuing (LLQ) is one of the newest Cisco queuing

mechanisms.

Congestion avoidance: Congestion avoidance techniques monitor network traffic loads.

The aim is to anticipate and avoid congestion at common network and internetwork

bottlenecks before it becomes a problem. These techniques provide preferential treatment

under congestion situations for premium (priority) class traffic, such as voice. At the same

time, these techniques maximize network throughput and capacity use and minimize packet

loss and delay. Weighted random early detection (WRED) is one of the QoS congestion

avoidance mechanisms used in Cisco IOS software.



6/10


Header compression: In the IP environment, voice is carried in Real-Time Transport

Protocol (RTP), which is carried in User Datagram Protocol (UDP), which is then put

inside an IP packet. This constitutes 40 bytes of RTP/UDP/IP header. This header size is

large when compared to the typical voice payload of 20 bytes. Compressed RTP (CRTP)

reduces the headers to 2 bytes in most cases, thus saving considerable bandwidth and

providing for better throughput.

Resource Reservation Protocol: Resource Reservation Protocol (RSVP) is a transport

layer protocol that enables a network to provide differentiated levels of service to specific

flows of data. Unlike routing protocols, RSVP is designed to manage flows of data ratherthan make decisions for each individual datagram. Data flows consist of discrete sessions

between specific source and destination machines. Hosts use RSVP to request a QoS level

from the network on behalf of an application data stream. Routers use RSVP to deliver QoS

requests to other routers along the paths of the data stream. After an RSVP reservation is

made, weighted fair queuing (WFQ) is the mechanism that actually delivers the queue

space at each device. Voice calls in the IP environment can request RSVP service to

provide guaranteed bandwidth for a voice call in a congested environment.

Fragmentation: Fragmentation defines the maximum size for a data packet and is used in

the voice environment to prevent excessive serialization delays. Serialization delay is the

time that it takes to actually place the bits onto an interface; for example, a 1500-byte

packet takes 187 ms to leave the router over a 64-kbps link. If a best-effort data packet of1500 bytes is sent, real-time voice packets are queued until the large data packet is

transmitted. This delay is unacceptable for voice traffic. However, if best-effort data

packets are fragmented into smaller pieces, they can be interleaved with real-time (voice)

packets. In this way, both voice and data packets can be carried together on low-speed links

without causing excessive delay to the real-time voice traffic.

There are many QoS tools that can be used to ensure consistent throughput. When these

mechanisms are employed, voice traffic on the network is assured priority and its delivery is

more consistent.


7/10

Reordering of Voice PacketsThis topic describes how RTP ensures consistent delivery order of voice packets in an

IP internetwork.


Reordering of Packets

IP assumes packet-ordering problems.

RTP reorders packets.

In traditional telephony networks, voice samples are carried in an orderly manner through the

use of time-division multiplexing (TDM). Because the path is circuit-switched, the path

between the source and destination is reserved for the duration of the call. All of the voice

samples stay in order as they are transmitted across the wire. Because IP provides

connectionless transport with the possibility of multiple paths between sites, voice packetscannot arrive out of order at the destination. Because voice rides in UDP/IP packets, there is no

automatic reordering of packets.

RTP provides end-to-end delivery services for data that require real-time support, such as

interactive voice and video. According to RFC 1889, the services provided by RTP include

payload-type identification, sequence numbering, time stamping, and delivery monitoring.

Example: Reordering Voice Packets

In the figure, RTP reorders the voice packets through the use of sequence numbers before

playing them out to the user.



8/10


The table illustrates the various stages of packet reordering by RTP.

Sequencing of Packets by RTP

Stage What Happens

Voice packets enter the network. IP assumes packet-ordering problems.

RTP reorders the voice packets. The voice packets are put in order through the use of sequence

numbers.

The voice packets are spaced according to the time stampcontained in each RTP header.

RTP retimes the voice packets.

The user hears the voice packets in order and with the sametiming as when the voice stream left the source.

RTCP (Real-Time TransportControl Protocol) sendsoccasional report packet fordelivery monitoring.

Both the sender and receiver send occasional report packetscontaining information, such as number of packets sent orreceived, the octet count, and the number of lost packets.


9/10

Reliability and AvailabilityThe traditional telephony network strives to provide 99.999 percent uptime to the user. This

corresponds to 5.25 minutes per year of downtime. Many data networks cannot make the same

claim. This topic describes methods that you can use to improve reliability and availability in

data networks.


Reliability and Availability

Traditional telephony networks claim 99.999%uptime.

Data networks must consider reliability andavailability requirements when incorporating voice.

Methods to improve reliability and availabilityinclude:

Redundant hardware

Redundant links

UPS

Proactive network management

To provide telephony users the sameor close to the samelevel of service as they experience

with traditional telephony, the reliability and availability of the data network takes on new

importance.

Reliability is a measure of how resilient a network can be. Efforts to ensure reliability may

include choosing hardware and software with a low mean time between failure, or installing

redundant hardware and links. Availability is a measure of how accessible the network is to the

users. When a user wants to make a call, for example, the network should be accessible to that

user at any time a call is required. Efforts to ensure availability may include installing proactive

network management to predict failures before they happen, and taking steps to correct

problems in design of the network as it grows.

When the data network goes down, it may not come back up for minutes or even hours. This

delay is unacceptable for telephony users. Local users with network equipment, such as voice-

enabled routers, gateways, or switches for IP Phones, now find that their connectivity isterminated. Administrators must, therefore, provide an uninterruptible power supply (UPS) to

these devices in addition to providing network availability. Previously, depending on the type

of connection the user had, they received their power directly from the telephone company

central office (CO) or through a UPS that was connected to their keyswitch or PBX in the event

of a power outage. Now the network devices must have protected power to continue to function

and provide power to the end devices.

Network reliability comes from incorporating redundancy into the network design. In

traditional telephony, switches have multiple redundant connections to other switches. If either



10/10


a link or a switch becomes unavailable, the telephone company can route the call in different

ways. This is why telephone companies can claim a high availability rate.

High availability encompasses many areas of the network. In a fully redundant network, the

following components need to be duplicated:

Servers and call managers

Access layer devices, such as LAN switches

Distribution layer devices, such as routers or multilayer switches

Core layer devices, such as multilayer switches

Interconnections, such as WAN links and public switched telephone network (PSTN)

gateways, even through different providers

Power supplies and UPSs

Example: Cisco Reliability and Availability

In some data networks, a high level of availability and reliability is not critical enough to

warrant financing the hardware and links required to provide complete redundancy. If voice is

layered onto the network, these requirements need to be revisited.

With Cisco Architecture for Voice, Video and Integrated Data (AVVID) technology, the use of

Cisco CallManager clusters provides a way to design redundant hardware in the event of Cisco

CallManager failure. When using gatekeepers, you can configure backup devices as secondary

gatekeepers in case the primary gatekeeper fails. You must also revisit the network

infrastructure. Redundant devices and Cisco IOS services, like Hot Standby Router Protocol

(HSRP), can provide high availability. For proactive network monitoring and trouble reporting,

a network management platform such as CiscoWorks2000 provides a high degree of

responsiveness to network issues.

2 1 requirements of voice

Documents